[00:38] <corpse> when i try to open my  web domain it opens up my router. i have turned off all firewalls and opened all ports and i still get the loop back. i also do not seem to beable to access it from outside of the network
[00:47] <abhibeckert> i'm setting up an Ubuntu 10.04 LAMP server, to be run inside a virtual machine on my workstation
[00:47] <abhibeckert> is it possible to give the VM 128MB of ram? I'd like to keep as much memory free as possible on the host machine
[00:47] <abhibeckert> It will only be used for apache/php/mysql and I'll be the only person it serves requests to
[00:49] <hggdh> abhibeckert: 128M may be a bit too low... but the best option is to try it, since you can always increase if needed
[00:50] <abhibeckert> thanks. i'll be following the slicehost ubuntu setup, which is full of config tips to get lamp running smoothly even with high traffic on their entry level 256MB VM's
[00:50] <abhibeckert> i'm hoping i can get it down a bit, since my workstation can be starved for ram... especially when i'm testing internet explorer in another virtual machine
[01:04] <qman__> abhibeckert, I run ubuntu server 8.04 on real hardware with 256MB, doing a whole lot more than LAMP
[01:04] <qman__> so I assume 128MB should work if you tweak it right
[01:21] <ruben23> hi what are the added features and goodness of the new version 10.04 LTS...?
[01:22] <therian> hey guys, can someone help me with ndiswrapper please?
[02:42] <ccheney> anyone happen to know how to have uec use lp logins, i can't remember what you have to do to make that work
[02:43] <ccheney> ah uec-run-instances -l (user id)
[03:00] <cef> argh, why does grub2 have stupid defaults for a server install? I want a menu damnit!
[03:04] <cef> ok, great.. 2.6.32-22-server breaks my kvm setup. grrrrrr
[03:05] <ccheney> cef, 22.35 ?
[03:11] <cef> probably.. got it from security today
[03:11] <cef> hold a sec I'll check (on an old kernel now -19-)
[03:13] <cef> yup, 2.6.32-22-server_2.6.32-22.35
[03:21] <lifeless> 35 is known broken, being fixed
[03:21] <Maletor> I'm trying to share my music through DAAP with RythmBox. Can I access this stream from work?
[03:24] <kees> cef: yes, known (bug 589163), fix will be published in a few hours.  :(
[03:27] <sludge321> Was wondering if anyone can help me with a strange issue on a new Ubuntu 10.04 server. /var/log/messages is empty, and cron.daily does not seem to be being processed.
[03:27] <ccheney> anyone happen to know what causes not tested instances when running uec-testing-scripts?
[03:29] <KurtKraut> sludge321, does the /var/log/messages modified date and time change? Is it being written and them wiped out? I'm clueless about what it might be, but certainly this is an important part of a diagnosis.
[03:30] <cef> kees: cool ta
[03:33] <a3ist> Ok so I've been drooling over the docs all day; can someone recommend a way to set up a new box w/ ubuntu server 10.04 and gdm/gnome with several simultaneous x sessions?  One will be for the HDTV its plugged into via HDMI, and the rest over the network.
[03:42] <sludge321> KurtKraut: looks like it isn't being modified, date on "ls -la" is 25th of May.
[03:44] <KurtKraut> sludge321, but is this partition mounted with an option that will update the 'last modified' that?
[03:44] <cef> also, is the partition mount rw?
[03:46] <sludge321> KurtKraut: cef: Hmm, "fdisk -l" results in "cannot open /proc/partitions" and "cat /etc/fstab" only has entried for /proc and /dev/pts
[03:46] <cef> sludge321: what does 'mount' say? it should list all the mounts
[03:48] <sludge321> cef: /dev/vzfs on / type reiserfs (rw,usrquota,grpquota)
[03:48] <sludge321> Maybe this is due to the type of virtualisation. I believe this server is a Parallels OpenVZ/virtuozzo vm.
[03:48] <Maletor> SSH seems slow from work. How can I speed up my ssh connection?
[03:49] <sludge321> kern.log is empty as well
[03:50] <sludge321> but mail and apache logs are being written to correctly
[03:50] <lucas_> mdadm reports that i have three spare disks in my system, but the system only has three disks total, how can i remove these from the set?
[03:55] <PcCowboy> I have a quick question. I'm trying to setup a home server for FTP, windows file sharing, HTTP, and any other cool servers i can come up with ;-P  I have baan using a windows XP box but i was thinking using ubuntu might mean it would run cooler (less heat), use less electric, ect..  My question is would server or desktop better suit my needs?
[03:55] <PcCowboy> *been
[03:56] <twb> lucas_: pastebin /proc/mdstat
[03:56] <twb> PcCowboy: http://mywiki.wooledge.org/FtpMustDie
[03:57] <twb> PcCowboy: if you want a server, install Ubuntu Server, not Ubuntu Desktop.
[03:58] <lucas_> pastebin gave a command not found, heres a cat
[03:58] <lucas_> root@hackintosh:/proc# cat mdstat
[03:58] <lucas_> Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
[03:58] <lucas_> md1 : active raid1 sdb2[2] sdc2[1]
[03:58] <lucas_>       1429225472 blocks [2/1] [_U]
[03:58] <lucas_>       [[03:58] <lucas_>       
[03:58] <lucas_> md0 : active raid1 sdb1[0] sdc1[1]
[03:58] <lucas_>       524287936 blocks [2/2] [UU]
[03:58] <lucas_>       
[03:58] <lucas_> unused devices: <none>
[03:58] <twb> !pastebin > lucas_
[03:59] <a3ist> Ok so I've been drooling over the docs all day; can someone recommend a way to set up a new box w/ ubuntu server 10.04 and gdm/gnome with several simultaneous x sessions?  One will be for the HDTV its plugged into via HDMI, and the rest over the lan
[04:00] <twb> a3ist: if by "LAN" you mean ethernet, that doesn't make sense.
[04:01] <twb> You can run X over TCP/IP over ethernet, but those are remote heads, not local heads.  So it'd be a normal terminal server plus a local head connected to your TV.
[04:01] <a3ist> got a new box that's doing media pc type stuff for the tv; gonna do mythTV or something there.  But I also want others on the lan to be able to log in and use a desktop environment without interrupting it
[04:01] <twb> Yeah, that's just a normal thin client/server setup.
[04:01] <lucas_> http://paste.ubuntu.com/444366/
[04:01] <twb> a3ist: will these other users be booting from the network, or will they be running (say) Windows and just opening some apps on your server?
[04:02] <a3ist> The latter.  Two mac osx boxes in my dad's home office
[04:02] <a3ist> I don't need DE's for my own use; I'm perfectly comfortable with CLI
[04:03] <twb> a3ist: OK, then then basically all you do is install (say) the ubuntu-desktop package for the TV head, and any apps you need, and set up openssh-server.
[04:03] <twb> a3ist: the OS X user runs (in X11.app) "ssh -X <server> <app> <app's args>"
[04:04] <twb> e.g. ssh -X mythtv firefox google.com
[04:04] <a3ist> I have gdm and the gnome desktop environment starting up fine, but I was looking for a way for my father to VNC in or something and get a normal GDM gui login.
[04:05] <a3ist> Right now it doesn't seem to let multiple gnome-sessions run at once
[04:05] <twb> a3ist: multiple sessions as the same user?
[04:05] <a3ist> no, different users
[04:05] <a3ist> each getting their own session
[04:05] <twb> It should.  I haven't tried it for a couple of years, but it should Just Work
[04:06] <twb> You can use VNC instead of X11 by installing vnc4server.
[04:06] <PcCowboy> twb: Im using an HTTP server too. whats a good way to do this on linux
[04:07] <twb> PcCowboy: have you read the Ubuntu Server Guide?
[04:07] <a3ist> I have x11vnc installed, but having multiple people use it just gets them the same display (:0)
[04:07] <twb> a3ist: that's because x11vnc is the wrong tool.
[04:07] <a3ist> ah, I'll try vnc4server
[04:07] <twb> a3ist: x11vnc exports an EXISTING display, like Windows VNC servers.  You want to export a NEW display, separate from the local head.
[04:07] <a3ist> right
[04:08] <a3ist> Ok, that would explain my problem then
[04:08] <twb> Windows VNC servers don't do that because Windows licenses (usually) only allow one user at a time.
[04:08] <PcCowboy> nope.just looking into thiss buddy...plus the only "info" i can find on server is cloud, Cloud, CLOUD, oh and more cloud....lol
[04:08] <twb> PcCowboy: install ubuntu-serverguide (or find it online) and read it.
[04:10] <PcCowboy> thanks ttyl
[04:12] <ScottK> The link to the server guide is in /topic.
[04:15]  * ccheney wonders if he has the test misconfigured or he is seeing some kind of bug
[04:17] <twb> ScottK: last time I looked that link was busted
[04:18] <ScottK> Works for me.
[04:18] <twb> I guess someone fixed it.
[04:28] <ccheney> kirkland, ping, you happen to around still?
[04:48] <panfist> I need to upgrade the nic on a server that I didn't set up. I will be replacing the onboard nic with an intel add-on
[04:48] <panfist> the nic was detected automatically, which is great, except it's named eth1 and the other is named eth0
[04:49] <panfist> is this difference relevant in any of the config files? or is it all config'd by IP?
[04:55] <cjs> Oooch. The 2.6.32-22 kernel upgrade killed all of my kvms.
[05:14] <kirkland> cjs: there's another update that fixes it
[05:14] <kirkland> cjs: sorry about that, we spent most of today unbreaking that
[05:15] <cjs> Is that update out yet? I just did the update a half hour ago, and discovered the problem.
[05:16] <cjs> I downgraded to confirm that it was the kernel upgrade, and it started working again, and I updated again just a few minutes ago but now I'm getting virt-manager complaining that it can't attach to dbus or something like that.
[05:16] <cjs> I suspect that that might be unrelated, but do you have any idea how I might fix that?
[05:17] <cjs> I don't appear to have a .gconf or .gconfd file in my home dir....
[05:18] <cjs> Ah, those were under root's home dir. Removing both .gconf and .gconfd entirely fixed the issue.
[05:19] <cjs> Anyway, so how do I get this new update that fixes this?
[05:25] <kirkland> cjs: should be out soon tonight/tomorrow
[05:26] <cjs> Got it, thanks. Will this then be -23?
[05:39] <qman__> panfist, all interface configuration is based on eth0, eth1, etc...
[05:39] <qman__> if you want the second NIC to replace the first one, you can change that setting in /etc/udev/rules.d/70-persistent-net.rules
[05:41] <cjs> Hm. How do I get a guest to boot single-user so that I can run a manual fsck?
[05:41] <cjs> It always seems to try to boot multi-user, tells me I need to run fsck manually, and then won't give me a shell prompt.
[05:54] <cjs> Well, I just booted an install CD and selected recovery mode.
[06:01] <hallyn> hmm - 'man lsof' gives a weird result in the options section
[06:28] <habanany> I don't get to get out of the black screen
[06:29] <habanany> Any Spanish speaker  ?
[06:29] <twb> !es
[06:31] <habanany> thanks ubottu
[06:32] <twb> habanany: try Ctrl+Alt+F1 (just a guess).
[06:32] <habanany> I will try twb , thanks
[06:36] <twb> I guess he was dual-booting
[06:42] <MasterZuFu> Hello everyone. I'm using ubuntu 10.04 for my server and am hosting a mybb forum site on it. For some reason captcha isn't working. I did some looking around and some people say I need something called "GD Library" or "AdEnhanced" installed on my server. Any idea what this is or how to install it?
[06:54] <MasterZuFu> nevermind, i figureed it out
[06:55] <abhibeckert> MasterZuFu: did you find a good tutorial? can you post it?
[06:56] <MasterZuFu> abhibeckert: http://www.redmezzanine.com/?p=360&lang=es
[06:57] <MasterZuFu> you just need these two commands: sudo apt-get install php5-gd
[06:57] <MasterZuFu> and service apache2 restart
[06:57] <MasterZuFu> depending on which version of apache ur running and whatnot
[06:58] <MasterZuFu> those commands might change depending on that, but that's the basic.
[07:09] <MasterZuFu> how do i block entire countries at once with a single command? (ip block)
[07:11] <abhibeckert> MasterZuFu: why would you do that?
[07:11] <MasterZuFu> because there's some countries I don't want to have access to my site.
[07:11] <abhibeckert> why?
[07:11] <MasterZuFu> why should it matter?
[07:12] <abhibeckert> only reasons i can think of are discrimination and security
[07:12] <abhibeckert> if it's security... i don't think it'd help any
[07:12] <MasterZuFu> it'd help a lot believe it or not
[07:12] <abhibeckert> how? IP blocks are easy to work around
[07:12] <MasterZuFu> blocking entire countries gets rid of a lot of spam and ddos attacks if
[07:13] <MasterZuFu> because no one's going to re-route their entire botnet through a proxy, then they'd ddos themselves
[07:13] <abhibeckert> it'd have a lot of false positives for spam, and just because you stop responding to ddos attempts doesn't mean they stop coming in and flooding your bandwidth
[07:13] <MasterZuFu> look. i don't have to justify why. i just asked how to do it.
[07:15] <abhibeckert> i've never done it, but i'm guessing you'll want to do it with iptables. that should help with googling at least
[07:15] <ttx> Good morning
[07:15] <MasterZuFu> morning/evening
[07:15] <MasterZuFu> someone said something about "cid" blocks or, something, i don't know exactly what it's called
[07:16] <MasterZuFu> but it allows you to block whole countries at once or something
[07:16] <abhibeckert> ip addresses aren't broken up by country
[07:16] <abhibeckert> they're broken up by ISP or something like that
[07:16] <MasterZuFu> i know
[07:16] <ajmitch> finding all addresses associated with a particular country would be tedious & error-prone because they're usually allocated in quick small non-contiguous blocks
[07:17] <MasterZuFu> i'm sure i'll find a way
[07:17] <ajmitch> especially for those allocated from certain places like APNIC
[07:17] <abhibeckert> so if you're trying to find all IP address ranges from, say, china... then it's probably going to be a huge list that changes all the time
[07:17] <MasterZuFu> yeah
[07:17] <abhibeckert> and it seems useless... since the USA is the second largest country sending out spam (as far as i know) and i'm assuming you don't want to block those?
[07:18] <MasterZuFu> spam isn't what i'm worried aobut
[07:18] <abhibeckert> maybe you'd be better off looking at what IP you're getting lots of traffic from, and blocking those
[07:18] <MasterZuFu> about**
[07:18] <abhibeckert> most likely there's only a few massive ISP's in china? you could find out what their IP blocks are
[07:18] <MasterZuFu> who mentioned china?
[07:19] <abhibeckert> no-one, i just gave it as an examlpe
[07:19] <MasterZuFu> ok. i'm just cautious as a chinese ip just tried to gain access to port 22 :)
[07:19] <abhibeckert> since you implied you're trying to block spam and ddos attacks... of which many come from chinese bots
[07:20] <MasterZuFu> which is why china is on my list to be blocked ;)
[07:20] <abhibeckert> i'm not a sysadmin... but i'm told it's a basically suicide to have port 22 on a public port
[07:20] <MasterZuFu> it's not public
[07:20] <MasterZuFu> only i can access it
[07:20] <MasterZuFu> but he was trying
[07:20] <abhibeckert> all of the servers run by our sysadmin either have a whitelist of a few IP's who can talk to 22, or run it on a random port
[07:21] <MasterZuFu> which is why i knew he was trying to access it
[07:21] <abhibeckert> do you have a fixed IP? maybe you could whitelist your IP and block everyone else's?
[07:21] <MasterZuFu> no i have dhcp
[07:21] <abhibeckert> it probably wasn't "he" who was trying to get into port 22, it was probably a botnet
[07:21] <twb> Disable password-based auth and use only key-based auth.
[07:22] <jefimenko> does anyone know of any util for automatically restarting services when they crash?
[07:22] <MasterZuFu> hmmmmmmmm didn't think of that
[07:22] <jefimenko> or is a cron hack the best practice?
[07:22] <abhibeckert> a lot of bots are in china, but not all of them. some people suggest most bots are in the USA
[07:22] <MasterZuFu> i know this
[07:22] <twb> jefimenko: put them in upstart
[07:22] <abhibeckert> anyway, you're best off making sensitive ports like 22 only available from specific IP's, perhaps find out what your ISP's allocated range is
[07:23] <abhibeckert> and then either use a crazy-ass long password, or use a private key
[07:23] <twb> abhibeckert: password length isn't relevant, since it'll hash down to a finite set
[07:23] <cef> I tend to run ssh on a non-standard port, not for security (the way I have it set up does that).. I do it to reduce the massive amount of logging from failed attempts that I would otherwise have on port 22
[07:23] <twb> cef: I would use hashlimit for that
[07:24] <abhibeckert> twb by "crazy-ass long" i just meant "too long to keep in your head"
[07:24] <cef> twb: I did.. still fills up
[07:24] <jefimenko> twb: looking at it now
[07:24] <MasterZuFu> my pass is already above 150 chars XD :P. I'm not familiar with ubuntu so I'm not entirely sure how to secure it. i did just pick up a copy of "Linux Server Security" published by O'reilly
[07:24] <twb> The big problem with using a nonstandard port is that most QoS heuristics won't bucket ssh traffic correctly
[07:24] <jefimenko> twb: it's installed by default in ubuntu?
[07:24] <twb> jefimenko: "it"?
[07:24] <jefimenko> twb: upstart
[07:25] <twb> Upstart has been Ubuntu's default (and only supported) init system for a LONG time.
[07:25] <abhibeckert> i gotta go guys, 5pm here
[07:25] <twb> Poor bastard must be a kiwi...
[07:25] <MasterZuFu> if it helps any i'm only an hour behind him XD
[07:25] <MasterZuFu> it's almost 4 pm here
[07:25] <ajmitch> twb: given that it's after 6pm in NZ..
[07:25] <cef> it's 4:30 here in melb
[07:26] <cef> (or close to)
[07:26] <ajmitch> 5pm, must be on a boat in the tasman
[07:26] <MJennings> Hi.. If i just installed ubuntu-server on a new pc and have two processes in D state that won't go away after reboot, what should i do?
[07:26] <MasterZuFu> nah, 5pm in australia in NSW
[07:27] <cef> his ip points him to qld
[07:27] <MasterZuFu> ooooo so close
[07:28] <cef> still only 4:30pm their time.. methinks he was confuzzled
[07:28] <MasterZuFu> ok, guys I gotta start reading this book or crazy chinese ppl are gonna hax my server XD
[07:30] <ajmitch> cef: easy when it's that close to beer o'clock on a friday
[07:30] <cef> ajmitch: heh
[07:34] <a3ist> anyone know exactly what the -query parameter of xvnc / xdmcp instances does?  Restricts it to that hostname?  All the examples are just for localhost.
[07:38] <cef> yay for kernel updates
[07:41] <cef> and there goes security.ubuntu.com's throughput as everyone else starts hitting it. ;)
[07:41] <kees> yeah, it should be on the -updates mirrors soon
[07:42] <cef> kees: yup.. it's up there now
[07:42] <kees> cool
[07:42]  * kees goes to bed now
[07:43] <cef> ahh hang on.. -updates.. yeah.. I think the aus mirror is doing an rsync.. it's complaining at me. ;)
[07:44] <kees> copying from -security to -updates usually takes an hour.  it just published on -security, so hopefully in an hour it'll be in -updates.
[07:44] <cef> :D
[07:48] <cef> yay! all good now
[07:51] <riz0n> hey guys, I have Ubuntu Server and a procmail filter system-wide that filters incoming mail flagged as spam to "Junk Mail" folder. I have a user receiving unwanted mail from an email address that wants to have mail from the email address moved to an IMAP folder (probably Junk Mail) can anyone here assist me in setting this filter up?
[07:57] <SpamapS> riz0n: if you're delivering to real system users, then adding the rule to ~user/.procmailrc will work just fine.
[07:57]  * SpamapS passes out
[07:58] <riz0n> yeah thats what im trying to do, but im not sure about how to write the rule.
[07:59] <riz0n> would this work? Line 1 :0:; Line 2 * ^From:.*\<email@ddresstofilter.com>; Line 3 Junk E-mail
[08:11] <jefimenko> twb: i already have an initscript for my service, but if it crashes, it doesn't come up
[08:11] <jefimenko> doesn't come back up
[08:12] <jefimenko> i've been reading the upstart docs and haven't found out yet how to auto restart a service that dies
[08:12] <jefimenko> if you know by experience how to do this, please point me into that direciton
[08:25] <ab2> anyone know how to change the default tty?  since installing xubuntu-desktop on my headless server i can no longer logon to ttyS0
[08:42] <twb> jefimenko: talk to #upstart about it
[08:42] <twb> I don't use it myself
[08:43] <MJennings> you have an option not to use upstart? how do you disable it
[08:44] <twb> By installing 8.04
[08:44] <twb> where upstart is really only a token implementation
[08:45] <ab2> how do you get vlans on top of bonding working using /etc/network/interfaces in 10.04?  i can't seem to find the proper configuration online.
[08:46] <MJennings> can xen be used on 10.04? i tried installing ubuntu-xen-server and there's dependencies problems
[08:47] <\sh> MJennings, you should read http://noone.org/blog/English/Computer/Debian/xen-tools%2520back%2520in%2520Debian%2520Unstable.html by axel beckert...xen-tools are missing in ubuntu and I think that's your dependency problem, right?
[08:48] <MJennings> \sh: correct
[08:48] <\sh> ab2, auto vlan<id> \n iface vlan<id> inet static\n <do the usual stuff> \n vlan_raw_device <your bond interface with out /dev/>
[08:49] <\sh> ab2, just like this: http://paste.ubuntu.com/444460/
[08:50] <\sh> MJennings, axel beckert took over upstream and debian maintainership for this package...as far as I understand he provides as well xen-tools package for ubuntu
[08:51] <\sh> ab2, the format is for lucid working with upstart
[08:57] <MJennings> \sh: what kind of name is axel beckert
[08:58] <\sh> mjeanson__, german ;)
[08:58] <ab2> thanks \sh =)  that worked.
[08:58] <\sh> ab2, np
[08:59] <lambrecht> morning
[09:21] <twb> b k e
[09:21] <RoyK> morning
[09:21] <RoyK> t w b
[09:22] <twb> I was trying to M-x s b k e
[09:22] <twb> Like so...
[10:23] <cybrocop> Hi All, I'm having problems with virsh. I accidentally did an init 0 while a kVM instance was running and after restarting my system I am no longer able to run any instances.
[10:23] <cybrocop> Virsh reports the instances as "started" but they don't produce any console output nor can I see anything with VNC (just a black screen).
[10:42] <TREllis> cybrocop: can you ping the ip's of the kvm instances ie are they up?
[10:42] <TREllis> there was a kernel bug affecting kvm recently (no instances started) https://bugs.launchpad.net/ubuntu/+source/linux/+bug/589163
[11:11] <ranjan> Hi all..how can i block google talk using squid proxy??
[11:13] <\sh> ranjan, google talk is xmpp...I wonder if they talk http / https at all
[11:13] <\sh> ranjan, or do you mean google talk on the google mail website?
[11:13] <ranjan> \sh, google talk on the google mail website  is the first priority
[11:14] <ranjan> \sh, how can i do that??
[11:14] <ranjan> \sh, i have a squid running here
[11:14] <\sh> ranjan, check the website for the javascript / html tag and try to block that...
[11:15] <ranjan> \sh, how..?? i want a solid solution. can you please help me?
[11:16] <cybrocop> TRellis: Thanks I think this is my problem.
[11:16] <\sh> ranjan, sorry..I can't help...I don't know anything about squid content filters...as I don't use gmail on the website
[11:16] <ranjan> \sh, oj
[11:17] <TREllis> cybrocop: should be a kernel upgrade available for it, it was a regression
[11:19] <cybrocop> Trellis: I may have done and apt-get upgrade to get a new kernel yesterday but I forgot to reboot. Right when I did an "init 0" (inadvertantly) the system was running several KVM images fine. As soon as I restarted, I started experiencing this problem. I'll check for updates.
[11:24] <TREllis> cybrocop: good idea
[11:48] <cybrocop> trellis: thanks, you saved me a lot of headache
[13:41] <twb`> I'm installing Lucid on a router, PXE-booting the d-i netboot kernel.
[13:41] <twb`> When it prompts for my "primary" interface, the eight ports are listed higgledy-piggledy.
[13:41] <twb`> Is there an easy way to tell which one is the "real" eth0?
[13:42] <twb`> http://pastebin.com/VAU1uVbG
[13:42] <twb`> Hm, never mind.  I just started brute-forcing them, and the first one worked (unlike the last boot).
[13:43] <ttx> twb`: heh
[13:51] <twb`> Argh, now it's falling over trying to wget the Release file, despite me being able to fetch it from the same URL from my own machine, with the same proxy settings.
[14:00] <sommer> morning :)
[14:01] <twb`> My own fault.  Apparently dhcp-range=vendor:PXEClient,[...] will still match DHCP requests from d-i :-/
[14:06] <ccheney> Daviey, i got a huge amount of not tested instances when i ran the test last night, and most of them according to the script were 'null', not sure if that is a bad thing or if it means the script itself is buggy
[14:07] <ttx> jdstrand: if you do some sync processing today, I'd appreciate if you could push tomcat6 (bug 582312, filed 2 weeks ago). I have a few work items blocking on this.
[14:09] <jdstrand> ttx: ok
[14:09] <ttx> jdstrand: thanks !
[14:11] <mathiaz> ttx: o./
[14:11] <ttx> mathiaz: yo
[14:11] <mathiaz> ttx: got some time for some hadoop talk?
[14:11] <ttx> mathiaz: sure, joining
[14:12] <ttx> mathiaz: btw you should still assign yourself to 1-2 bugs in https://launchpad.net/server-papercuts/+milestone/maverick-alpha-2
[14:12] <mathiaz> ttx: yeah - on my TODO list
[14:12] <ttx> I intend to push ivoks on bug 321091, fwiw
[14:14] <twb`> Is there *ANY* downside to installing ext3 and then remounting as ext4 at some later date?
[14:14] <ccheney> mathiaz, what causes uec-testing-scripts to return 'null' for not-tested instances? i ran a 1000 run test last night and 721 of them showed up as not-tested 'null'
[14:15] <ccheney> twb`, you don't get all the features unless it is formatted as ext4
[14:15] <ccheney> twb`, but i don't happen to remember the details
[14:15] <twb`> "all the features" boils down to some minor optimizations, plus extents.
[14:16] <ccheney> twb`, probably something like that, yes
[14:16] <twb`> And if I enable extents, I lose the ability to mount as ext3, which has screwed me a few times in the past
[14:16] <ccheney> ah, so perhaps its not a feature in your case anyway :)
[14:16] <twb`> I'm not saying extents aren't useful
[14:17] <twb`> (I'm installing what will be our new mission-critical router.)
[14:18] <twb`> Actually, I'm installing onto a CF card, so ext*2* is almost attractive.
[14:18] <Daviey> ccheney: saw your email.. i don't know tbh.. hggdh might be a better person to ask
[14:19] <ccheney> Daviey, ok
[14:19] <ccheney> hggdh, ping
[14:24] <twb`> I picked ext3 for now, since I can easily migrate to ext2 or ext4 at a later date.
[14:32] <rgreening> ScottK: hey. how doth bug 589188 fare?
[14:33] <ScottK> rgreening: Approved.  Just waiting for an archive admin to execute it (if you ask Riddell nicely ...)
[14:33] <rgreening> cool
[14:39] <twb`> When doing a priorty=low d-i install, linux-generic is i686, and linux-386 is for i386?
[14:41] <\sh> ScottK, any clue on "when do sync requests be processed?"
[14:41] <ScottK> \sh: In theory every day by the archive admin of the day.  In practice it seems like on Tuesday's when it's Riddell's day.
[14:41] <\sh> oh well...
[14:43] <twb`> Paraphrasing: what differentiates linux-image-386 from linux-image-generic?
[14:43]  * twb` digs out their .configs...
[14:52] <twb`> Apparently there's bugger-all difference
[14:52] <twb`> http://pastebin.com/SuQJkV7d
[14:53] <wise_crypt> is it possible to copy squid cache file from other squid server ?
[14:53] <twb`> wise_crypt: try #squid.
[14:54] <twb`> It'll depend on the internal format squid users
[14:54] <twb`> s/users/uses/
[14:55] <wise_crypt> twb`: ex. from an oss squid server to ubuntu squid server
[14:56] <twb`> wise_crypt: oss?
[14:57] <wise_crypt> open suse
[14:58] <twb`> Urk.
[14:58] <twb`> SLES makes *RHEL* look good.
[15:01] <wise_crypt> twb`: opensuse and it uses amd processor
[15:02] <wise_crypt> twb`: the ubuntu uses intel, can it be done ?
[15:02] <sjm> wise_crypt: why do you need to?  The cache should regenerate on its own
[15:03] <twb`> wise_crypt: squid is extremely unlikely to give a shit what CPU you run
[15:04] <wise_crypt> sjm: it tooks time to generate, a couple of moths ahead with high traffic will have plenty of resources
[15:04] <twb`> So this whizzo encryption of /home -- will it DTRT if deployed on a multi-user CIFS/NFS NAS?
[15:05] <sjm> wise_crypt: and you can use wget with the --delete-after option to help generate it if you know what pages you want in there.
[15:05] <twb`> sjm: it'll still involve a second request to the internet, unless he points one squid at the other
[15:06] <sjm> wise_crypt: run the wget jobs overnight (or off-peak)
[15:06] <twb`> Actually... you can tell squid about "peer" caches.  If the old SUSE junk will continue running, you could just leave it there.
[15:06] <sjm> wise_crypt: what twb said.  use the SUSE cache as a parent of the new one.
[15:09] <wise_crypt> sjm twb` to do that i have to buy a new server :) unfortunately the budget is too low for this
[15:10] <twb`> Yeah, I thought he might be reflashing
[15:11] <wise_crypt> migrating is always pain
[15:11] <twb`> Migrating from SUSE is, sure
[15:11] <wise_crypt> :))
[15:12] <wise_crypt> the hell let them bugging
[15:13] <wise_crypt> *the hell let them bugging me with me the connection (solved)
[15:13] <wise_crypt> thanks twb` sjm
[15:32] <mathiaz> ccheney: hm - not-tested returns null - I don't know what can leads to that
[15:32] <mathiaz> ccheney: do you have a log of the test run somewhere?
[15:40] <ccheney> mathiaz, yes
[15:40] <ccheney> mathiaz, will email you the log, it doesn't have debug info in it but if needed i can probably generate one with debug with nulls too
[15:40] <mathiaz> ccheney: coudl you put on people.canonical.com instead of emailing it?
[15:40] <ccheney> ok
[15:41] <ccheney> mathiaz, http://people.canonical.com/~ccheney/ccheney.log.2010-06-04_010115
[15:43] <mathiaz> ccheney: null probably comes from the fact the instance didn't have an Instance id set
[15:43] <mathiaz> ccheney: looking at the log: WARNING:root:Instance failed to start: FinishedVerify: Not enough resources available: addresses (try --addressing private)
[15:43] <mathiaz> ccheney: ^^ so it seems that instances were not started at all (thus not instance id was available)
[15:44] <ccheney> oh ok
[15:44] <ccheney> hmm
[15:44] <ccheney> how many ips do i need to have available for it to not fall over when testing like that, i have 2 threads and 16 ips
[15:44] <ccheney> i noticed that before but didn't realize that was what caused all the nulls
[15:45] <ccheney> when i looked at the running box i noticed it seemed to get to the end of the pool and never reuse the ips that were no longer in use
[15:45] <mathiaz> ccheney: hm - good question - I don't know exactly
[15:45] <mathiaz> ccheney: the script will not start instances if there aren't enough ressource available
[15:46] <mathiaz> ccheney: however it seems that UEC may take some time to give back public ip addresses
[15:46] <mathiaz> ccheney: you may have run into that issue where there was enough ressource available but no public ip were available yet
[15:46] <ccheney> yea and it we are going through an ip every 30 seconds when testing (seems that way on my laptop anyway) then it could use quite a lot
[15:46] <Nicd^> I just installed Ubuntu Server 10.04 and cannot get gnome to install, "Depends: swfdec-mozilla but it is not going to be installed"
[15:47] <mathiaz> ccheney: IIRC it should check if there are enough ressource available every 30 seconds or so
[15:48] <mathiaz> ccheney: 16 ips should be enough though - if you just have two cores on your laptop
[15:48] <ccheney> yea just a 2 core laptop
[15:48] <mathiaz> ccheney: 2 cores means you can run up to 2 instances at the same time
[15:48] <mathiaz> ccheney: so something else is fishy here
[15:48] <ccheney> yea
[15:49] <mathiaz> ccheney: you may have encountered an UEC bug then
[15:49] <ccheney> i may have done something wrong or found a bug
[15:49] <ccheney> i'm new to UEC, but i followed the directions as far as i can tell :)
[15:49] <mathiaz> ccheney: rerun the test script with debug mode
[15:49] <ccheney> mathiaz, ok
[15:49] <mathiaz> ccheney: is this running on maverick?
[15:50]  * Daviey laughs at mathiaz for suggesting any success on maerick
[15:50] <Daviey> maverick*
[15:50] <ccheney> mathiaz, lucid all updates but the 22.35 kernel
[15:51] <ccheney> and with my new sru roll
[15:52] <mathiaz> ccheney: I'd try to start a couple of instances by hand
[15:52] <mathiaz> ccheney: I don't think it's an issue with the script
[15:52] <ccheney> a small 20 debug run probably will show up the nulls i'll try that first
[15:52] <ccheney> mathiaz, yea its been working all this time, i thought maybe i screwed it up somehow, heh
[15:54] <twb`> Hm, git is still git-core in lucid
[15:54] <twb`> I thought it post-dated the removal of the old git package
[15:55] <ccheney> WTF
[15:55] <ccheney> its still saying i'm out of ip addresses
[16:00] <ccheney> oops, forgot to run the test as debug, grr :-\
[16:00] <ccheney> i did get 80% failure on a 20 instance test though which seems rather odd
[16:01] <ccheney> er not failure, -> not tested
[16:01] <twb`> Wow, sensible-editor(1) prompts you to pick an editor now?
[16:03] <hggdh> ccheney: pong (sorry, but had to get a voters registration done)
[16:04] <ccheney> hggdh, ok np, mathiaz is looking at my uec-testing-scripts logs to see why its outputting lots of not-tested 'null' items
[16:04] <hggdh> ccheney, mathiaz: these are sessions that failed even to start
[16:04] <ccheney> hggdh, appears to be, seems to possibly be an issue with running out of ips
[16:05] <ccheney> hggdh, it seems it takes a long time before they are returned to the pool after an instance termination
[16:05] <hggdh> ccheney: yes, make sense
[16:05] <hggdh> ccheney: I was told that "eventually they return to the pool"
[16:06] <ccheney> heh and when doing load testing eventually can be long enough to cause not-tested
[16:06] <hggdh> ccheney: most certainly
[16:06] <ccheney> i probably need 32 ips for a 2 thread system from a hunch i have
[16:07] <ccheney> at least if they are actually released after they no longer show up in euca-describe-instances
[16:07] <ccheney> however i was having problems with no ips even when there weren't many in that list, so i think it might not be related to that
[16:08] <hggdh> ccheney: I think it goes beyond that, yes
[16:08] <habanany> which one u recommend eBox, ISPConfig or Webmin ?
[16:09] <cloakable> I recommend /bin/bash :P
[16:09] <pmatulis> cloakable: what about /bin/zsh?  :)
[16:10] <cloakable> pmatulis: get thee from my sight, heretic ;)
[16:10] <twb`> zsh is for csh refugees
[16:10] <pmatulis> the zee shell!
[16:10] <cloakable> bash is for people who like to get work done ;)
[16:11] <twb`> The zsh author chose the name because it collided with one of his lecturers.  That shows how unconscionably rude he is.
[16:11] <twb`> *collideded with one of his lecturer's names, that is
[16:11] <cloakable> heh
[16:13] <twb`> I would say: bash is for people who have more important things to do, when being told to fix a new server, than installing another shell.
[16:13] <cloakable> pmatulis: next you'll be suggesting that emacs be put into ubuntu's /etc/shells ;)
[16:14] <twb`> cloakable: that breaks scp/sftp; emacs -c does The Wrong thing.
[16:14] <twb`> It works OK as init, though.
[16:14] <ccheney> hmm it seems like euca ate my ips but one
[16:14] <ccheney> er ate all
[16:15] <cloakable> twb`: file a bug report ;)
[16:15] <twb`> Incidentally, tmux got that right.
[16:15] <twb`> "tmux -c foo" will just run foo in a "real" shell.
[16:16] <cloakable> heh
[16:21] <ewait> anyone know about channel scanning for a backend non gui server?
[16:21] <twb`> ewait: as in 802.11?
[16:21] <ewait> no, as in atsc or dvb-c (tv channels)
[16:22] <twb`> No idea.
[16:22] <ewait> thx anyway
[16:22] <twb`> Try looking at the packages that depend upon <the dvb library package>
[16:23] <ewait> I have dvb-util and dvb-apps but they seem like there is no US channel presets
[16:27] <ccheney> mathiaz, the other two are in the same spot now
[16:28] <ccheney> mathiaz, of the two new runs the first one was without debug by accident and then ran the second one after it
[16:30] <ccheney> mathiaz, it seems to be stuck only using 10.0.0.46, my pool is 10.0.0.32-10.0.0.48
[16:30] <osmosis> i did a fresh install of ubuntu server 10.04 64bit. Im getting super slow disk access times. Not sure why.
[16:30] <osmosis> any ideas on how I can test / troubleshoot?
[16:30] <twb`> iostat/iowait
[16:30] <bcessa> hi, I'm trying to set nginx as a proxy to handle static files while let apache proccess the php code, how ever I'm getting a 502 Bad Gateway error, can someone give a hand with it?
[16:31] <twb`> bcessa: examine the entire response with curl -v.
[16:31] <osmosis> twb`, whats that?
[16:31] <habanany> ebox vs webmin , help me to decide
[16:32] <twb`> iostat               (1)  - Report Central Processing Unit (CPU) statistics and input/output statistics for devices and partitions
[16:32] <twb`> hdparm -i is another obvious one -- check it's using DMA
[16:32] <bcessa> twb`: this is what I get back http://pastebin.com/vbjcFPFY
[16:33]  * ccheney going to reboot his cloud and see if it passes after doing that
[16:33] <twb`> bcessa: ask the nginx people; I don't know.
[16:34] <osmosis> twb`, dma modes look okay. I have two brand new western digital drives, ubuntu 10.04 install with software raid1 and LVM
[16:34]  * ccheney thinks the not-tested are a combination bug/feature, in that if the ips are all used and you get to 12 attempts to get an ip it probably permanently ignores the instance
[16:34] <twb`> osmosis: are the partitions/filesystems aligned on block boundaries?
[16:34] <bcessa> I'm already on their forum so far nobody is answering :O thnx anyway
[16:34] <twb`> though that may constitute ricing...
[16:34] <osmosis> twb`, uhmm...it was all through the ubuntu installer.
[16:34] <ccheney> but there seems to also be a bug on my box where all but one ip is never released
[16:34] <twb`> osmosis: then probably not
[16:35] <twb`> osmosis: ref. http://thunk.org/tytso/blog/2009/02/20/aligning-filesystems-to-an-ssds-erase-block-size/
[16:35] <osmosis> twb`, these are not SSD drives. just normal drives.
[16:35] <twb`> Are they 2TB drives?
[16:36] <osmosis> twb`, yes
[16:37] <twb`> Still largely, then applicable.
[16:37] <bcessa> is weird, the problem is specifically with php files, plain html files works fine, example: http://ve.6pprv6tr.vesrv.com/index.html
[16:37] <twb`> Urk
[16:37] <twb`> *still largely applicable, then
[16:37] <bcessa> and this don't work: http://ve.6pprv6tr.vesrv.com/test.php
[16:38] <osmosis> twb`, i dont see any CPU slowing on my munin charts though. if it was overhead, it should be high iowait.
[16:38] <twb`> osmosis: block misalignment will result in IO waiting, not in CPU load
[16:39] <twb`> Anyway, if this is a fresh install, how do you know it's slower than it should be?
[16:40] <osmosis> twb`, because i have a slower server that was running the same load just fine. this is unusable.
[16:40] <osmosis> twb`, i dont have any graphs showing a lag...but the server is often unresponsive
[16:40] <osmosis> twb`, actually i do see a lot of iowait on the munin cpu chart.
[16:41] <twb`> Define "slower"
[16:41] <osmosis> twb`, tcp connections being dropped, slow apache load time, and i have had disk corruption as well
[16:42] <osmosis> twb`, its definitely IOWAIT
[16:42] <osmosis> twb`, the article you sent is just talking about SSDs though. not sure i understand it yet
[16:43] <bcessa> I notice that the php request is not being redirected to apache, and instead handle directly by nginx, so that what's causing the bad gateway, now the question is why :O
[16:43] <twb`> disk corruption isn't the same as "it's slow"
[16:43] <twb`> That indicates that something is really, seriously wrong.
[16:43] <osmosis> twb`, yes
[16:43] <edwin-sv> hi everybody, someone know about a package or related procedures with gnu/linux commands to make an "image" of an ubuntu server that is running but it must not being shutdown to make an image (maybe with Nort*nGhost or related software) because it is providing a critical web service. The purpose of the task I want to do is to keep a backup image of the server that actually it does not exists. :(
[16:43] <twb`> osmosis: you should indicate that first.
[16:43] <twb`> s/indicate/investigate/
[16:44] <twb`> OK, I'm obviously too tired and drunk to be helping
[16:44] <osmosis> twb`, investigate? i dont have much to go on. cant reproduce it
[16:44] <edwin-sv> thanx in advance! :)
[16:44] <osmosis> twb`, but my IOWAIT is high, and thats my best lead.
[16:48] <ccheney> on a fresh rebooted setup it seems to be working ok, at least for a while
[16:48] <ccheney> i might need to have it run for a few hundred instances with debug logging
[16:49] <ewait> how many things will break when doing a sudo do-release-upgrade?
[17:03] <bcessa> cya
[17:03] <ewait> My concern is with backuppc, virtual hosts in Apache2, mysql, etc. with the release upgrade
[17:06] <edwin-sv> hi everybody, someone know about a package or related procedures with gnu/linux commands to make an "image" of an ubuntu server that is running but it must not being shutdown to make an image (maybe with Nort*nGhost or related software) because it is providing a critical web service. The purpose of the task I want to do is to keep a backup image of the server that actually it does not exists. :( < Thanks in advance! :-) >
[17:07] <ccheney> no nulls on a rebooted box with 32 run, trying 256 run without rebooting first
[17:13] <oru_work> which program from repos can I install to be able to traceroute ?
[17:18] <eagles0513875> !kvm
[17:18] <ccheney> edwin-sv, a fs snapshot might work, but even then it might leave it in inconsistent state for apps
[17:18] <ccheney> edwin-sv, i think there are some tools to do backups utilizing snapshots
[17:19] <ccheney> edwin-sv, but i don't have any experience with them
[17:21] <ccheney> its not failing at the moment but it definitely looks like it might be eating ips
[17:21] <ccheney> at least based on how its reusing them
[17:21] <edwin-sv> ccheney, that is the situation, actually i have an ubuntu server providing a critical service to our enterprise, but it doesn't have a backup of configurations of its apps
[17:22] <ccheney> edwin-sv, this might be useful, but i am not sure if there is already something in ubuntu that makes it easier: http://tldp.org/HOWTO/LVM-HOWTO/snapshots_backup.html
[17:24] <osmosis> im still not sure why im getting such high IOWAIT on these new WD 2TB drives.
[17:24] <xperia> hello to all. i have a question related to uploading stuff to the server. why does i get allways a 405 Error Message "The requested method PUT is not allowed for the URL" everytime i try to upload something as follow "curl -T file.gz http://serverip/dir/file.gz"
[17:24] <edwin-sv> ccheney, do you know if a fs snapshot could take an image of the boot procedure, too?
[17:25] <xperia> do i need to change some server settings as ubuntu prevent this and if yes where do i need to change this ?
[17:25] <ccheney> edwin-sv, you would probably have to reinstall grub, but if the (i assume raid) complexity is hidden from the OS then it probably would work otherwise
[17:25] <ccheney> edwin-sv, i have no idea what you would need to do to restore if you are using softraid
[17:26]  * ccheney is a newbie server guy so doesn't know detail about a lot of these things
[17:27] <ccheney> hmm yep euca is eating my ips at least its doing it somewhat slowly
[17:27] <osmosis> I have two 2TB western digital drives, on software raid1 and LVM. Im getting high IOWAIT. How can I determine the cause?
[17:31] <xperia> anybody know how i can resolve this Problem here "The requested method PUT is not allowed for the URL"  ?
[17:32] <guntbert> xperia: for apache problems there is the channel #httpd too
[17:35] <xperia> guntbert: but nobody answer there. tryed allready !
[17:36] <edwin-sv> ccheney, actually I want to have an image of the ubuntu server that is running, and if fail the one that is running just turn on the image in other server the image and keep running the apps
[17:36] <guntbert> xperia: I cannot help you either - sorry - it was only meant as additional ressource :-)
[17:37] <edwin-sv> ccheney, actually I want to have an image of the ubuntu server that is running, and if fail the one that is running just turn on the image in other and keep running the apps :-)
[17:37] <hggdh> smoser, mathiaz, ccheney: can I reboot the UEC test rig?
[17:38] <mathiaz> hggdh: sure
[17:38] <hggdh> thanks
[17:39] <RoyK> hm. ubuntufying zimbra would be a nice project
[17:39] <ccheney> hggdh, yep thats fine, i'm running uec testing on my local setup
[17:41] <hggdh> ccheney: thank you
[17:42] <ccheney> mathiaz, did you happen to see my debug log i posted earlier?
[17:42] <mathiaz> ccheney: not yet
[17:42] <mathiaz> ccheney: could you repost the url?
[17:42] <ccheney> mathiaz, http://people.canonical.com/~ccheney/
[17:42] <ccheney> mathiaz, the latest timestamp is the full debug log
[17:43] <ccheney> mathiaz, its looking like eucalyptus will reuse an ip immediately until some trigger and then either never reuses it or takes a very long time to do so, which appears to be why i was running out of ips, not sure if that was the cause of the 'null's though
[17:49] <mathiaz> ccheney: well - the script is running as expected IMO
[17:49] <ccheney> ok
[17:49] <ccheney> so a null just refers to something that never got an ip address?
[17:49] <mathiaz> ccheney: the second time an instance is trying to run, uec fails: DEBUG:root:Instance start output: FinishedVerify: Not enough resources available: addresses (try --addressing private)
[17:50] <ccheney> yea
[17:50] <mathiaz> ccheney: so null refers to the fact that an Instance object has been created in but it failed to start (which means that no instance Id are unkown for the object)
[17:50] <mathiaz> ccheney: that's why there is a null in the report
[17:50] <ccheney> oh i see
[17:51] <mathiaz> ccheney: hm - no instance Id are *known* to the testing script
[17:51] <ccheney> yea
[17:51] <mathiaz> ccheney: that's why it fails to null
[17:51] <ccheney> ok
[17:51] <hggdh> and which is why is I left it as 'not-tested'
[17:51] <mathiaz> ccheney: the script reveals a problem with UEC though
[17:51] <mathiaz> ccheney: not a real bug in the script ;)
[17:52] <ccheney> mathiaz, ok
[17:54] <ccheney> i'm at the last ip on a debug log 256 run, so maybe it will be of some help to determine whats going wrong
[17:55] <ccheney> hmm i think i'll reboot so its the full log from a clean system
[17:55] <ccheney> it got to the problem in under 30m
[18:08] <ccheney> i see this in the cc.log is that bad: [Fri Jun  4 12:05:35 2010][001251][EUCAWARN  ] vnetInitTunnels(): in MANAGED-NOVLAN mode, priv interface 'eth0' must be a bridge, tunneling disabled
[18:16] <Tim_R> can anybody tell me how to setup quota for webspace because I want to offer my customers 25mb free webspace
[18:16] <g-hennux> hi!
[18:16] <g-hennux> from 9.10 to 10.04, apparently the dovecot-config file has changed
[18:17] <g-hennux> and i denied the upgrade of the config file at dist-upgrade
[18:17] <g-hennux> how can i get the default config file from the deb file to compare the two files manually?
[18:17] <ccheney> hmm lease time for euca dhcp is 20m if its not shut down properly
[18:17] <coafcv> how can I override the user password? I have a VM running ubuntu server, but I forgot both username and password. it's on my local LAN, and re-installing the OS would take too long.
[18:17] <ccheney> er 20m, which is  maximum time of lease if its not shut down properly is more accurate
[18:19] <Daviey> ccheney: yeah, so 20m suggests a bigger pool would hide that concern
[18:19] <Daviey> ccheney: did you recieve kirklands mail?
[18:19] <Tim_R> anybody?
[18:20] <Daviey> If you have a branch or debdiff you want me to have a gander over, let me know (via email ideally)
[18:20] <ccheney> Daviey, just saw it yea
[18:20] <Daviey> Tim_R: i would first point you towards disk quota
[18:20]  * Daviey goes afk
[18:20] <ccheney> Daviey, now that i know what to look for i am going to make sure the problem is really just the lease time and then bump my pool size
[18:20] <Tim_R> ok Daviey afterwares
[18:22] <ccheney> i have a doctor appointment (should be ~ 15m) and a house closing (~ 30m) today also, as noted on my calendar
[18:25] <ccheney> Daviey, i'm not sure but it doesn't look like the 'public' addresses are assigned by dhcp
[18:26] <ccheney> Daviey, and from what i recall when you log into a instance you see the internal ip eg 172.x
[18:26] <ccheney> it might still be related to the 20m lease time for the internal 172.x addresses but not certain, i'll get a good debug log and then up the number of ips
[18:34] <mathiaz> kees: hi - do you know if there is a package that provides the Debian CA certificate?
[18:34] <mathiaz> kees: I'm trying to connect to alioth over ssl and firefox doesn't recognize the CA
[18:35] <kees> mathiaz: hm
[18:36] <kees> mathiaz: I thought ca-certificates came with it
[18:36] <mathiaz> kees: http://ca.debian.org/ -> says it's from SPI
[18:37] <kees> http://www.spi-inc.org/secretary/spi-cacert.fingerprint.txt/view?searchterm=cert
[18:37] <kees> or rather: http://www.spi-inc.org/secretary/spi-cacert.crt/view?searchterm=cert
[18:39]  * ccheney thinks he found out his problem but isn't quite sure what the minimum ip range is yet, still reading a comment from developer about a similar user issue
[18:44] <ccheney> i think maybe i have to tweak VNET_ADDRSPERNET=32 to something smaller
[18:44] <elysrum> Hi, looking for a little help here.  Every time my server reboots something is resetting the permissions on /tmp to 755.  Anybody have any ideas?
[18:49] <elysrum> Server Version is 10.04, it is an upgrade from Karmic (not a clean install).
[18:54] <hggdh> kirkland: I wonder if Deji's systems really have KVM disabled, or if the KVM mods were not loaded -- just had the same issue on sapodilla
[18:56] <hggdh> Deji1: ping
[18:57] <Deji1> hggdh : pong
[18:59] <ccheney> actually it seems that was not the issue as the user was seeing their dhcp show up with all 0's
[19:01] <hggdh> Deji: did you reboot your NC?
[19:02] <Deji> hggdh : i have pricklypear up but still no luck with ceylon
[19:04] <hggdh> Deji: can you please, on ceylon, run 'initctl status qemu-kvm' and tell me what is the output?
[19:08] <Deji> hggdh : the system is down now, i can't even ssh into it. port 22 is not active
[19:10] <hggdh> Deji: OK. We wait for it to come back on, and then we check again
[19:15] <nick58b> uptime
[19:25] <a3ist> hey guys I'm looking for the current config method for xdmcp; I'm not seeing gdm.conf anywhere and I'm assuming it was moved to an auto-generation setup like grub2.  Need to know where to go enable the gdm greeter for remote logins.
[19:27] <a3ist> nm think I found it
[19:41] <Doonz> so my keyboard mapping is all fucked up . where can i change these settings. 10.4 console only no desktop
[19:47] <alvin> My fileserver just died. PSU was broken. I replaced it and booted. It has 4x2TB disks in mdadm RAID5. Now I'm staring at a big purple screen saying" Your disk drives are being checked for errors, this may take some time". I have the option to press C to skip. Now my question is: is this a rebuilt, or an fsck?
[19:48] <alvin> What I really want now is some output. A message saying fsck, or the ability to look at /proc/mdam. Something more informative than a purple background.
[19:48] <RoyK> alvin: reboot into single and check
[19:49] <alvin> But it's checking. I don't know if it is safe to reboot
[19:49] <RoyK> press escape
[19:49] <alvin> aha!
[19:49] <RoyK> that usually takes it back to the console
[19:50] <alvin> mountall terminated with status 1 on some filesystems. And ureadahead-other terminated with status 4. Those are the last lines
[19:50] <alvin> thanks btw
[19:53] <alvin> ahm the ureadahead message is a bug. It's now continuing. This is pretty fast for an fsck. Must be the ext4
[19:59] <alvin> Why extreme NFS timeouts are great: the fileserver was hosting the qcow image of a virtual mailserver. The mailserver is accessible again and I didn't need to reboot it while it's disk was offline.
[20:00] <alvin> Must also be a KVM thing, because my desktop on NFS mounted /home is still unresponsive
[20:03] <mathiaz> ccheney: hey - what's the state of the eucalyptus sru
[20:03] <mathiaz> ccheney: ?
[20:03] <ccheney> mathiaz, sent the debdiff to dave and dustin and was trying to do initial testing when i started getting all those weird errors
[20:10] <Deeps> can anyone recommend any hackable web interfaces for iptables?
[20:12] <Deeps> or possibly suggest a better solution to my conundrum
[20:12] <Deeps> i have an ubuntu router for my connectivity
[20:12] <Deeps> the router has multiple upstreams that can be used
[20:13] <Deeps> at present, i'm using source based routing using fwmarks to match
[20:13] <Deeps> e.g., for i in {2..10}; do ip rule add fwmark $i lookup $i; done
[20:14] <ccheney> mathiaz, would you like a copy also?
[20:14] <mathiaz> ccheney: yes
[20:14] <Deeps> and then mangle rules to match specifiuc client IPs
[20:14] <Deeps> e.g. iptables -t mangle -I PREROUTING -s 192.168.1.3 -j MARK --set-mark 3
[20:14] <Deeps> what i'd like is an easy way for client machines to advise the router what upstream they'd like to use
[20:15] <Deeps> so my idea was to have a web interface that would alter the mangle table rules
[20:15] <Deeps> and have that run on / interface with the router
[20:17] <ccheney> mathiaz, ok sent, let me know if you get it
[20:46] <Deeps> nobody? heh
[20:54] <EtienneG> question, guys: can you define mutliple static per interfaces in /etc/network/interfaces without using stupid eth0:1 type of alias?
[20:55] <Deeps> EtienneG: up/post-up lines with ip address add $ip/$mask dev $interface
[20:55] <EtienneG> Deeps, ugh.  thanks for the info, though!
[20:55] <Deeps> remember to add the corresponding down lines too though
[20:56] <Deeps> if you're wanting to add a range of ips, post-up for i in {1..255}; do ip addr add blah.$i/$mask dev $int; done
[21:04] <mathiaz> ccheney: ok - I got your diff
[21:04] <mathiaz> ccheney: could you create a bzr branch instead
[21:04] <mathiaz> ccheney: and create a merge proposal
[21:04] <mathiaz> ccheney: it would easier to conduct the review
[21:04] <ccheney> ok
[21:04] <mathiaz> ccheney: and subscribe other - rather than just sending debdiff around
[21:05] <mathiaz> ccheney: bzr co lp:ubuntu/lucid-proposed/eucalyptus
[21:05] <ccheney> ok
[21:05] <mathiaz> ccheney: have you used bzr pkg branches already?
[21:06] <ccheney> no, never did a merge proposal, etc
[21:07] <mathiaz> ccheney: ok - let me guide through this
[21:07] <mathiaz> SpamapS: around?
[21:07] <ccheney> i have mostly just worked on OOo and some internal oem stuff in the past which didn't use bzr branches
[21:07] <mathiaz> ccheney: right
[21:07] <ccheney> i'm going to have to leave in about 20-30m to go to a home closing (sale)
[21:07] <mathiaz> SpamapS: I was about to guide ccheney through using bzr pkg branches for doing an SRU
[21:08] <ccheney> but if it doesn't take much time i can do it now :)
[21:08] <mathiaz> ccheney: ok - I can get you started on this
[21:08] <mathiaz> ccheney: I usually start by creating a shared repository to hold all the branches:
[21:08] <mathiaz> bzr init-repo eucalyptus
[21:08] <mathiaz> naming the directory after the source package
[21:09] <mathiaz> then cd eucalyptus/
[21:09] <ccheney> ok
[21:09] <mathiaz> bzr co lp:ubuntu/lucid-proposed/eucalyptus
[21:09] <mathiaz> ^^ that gets the code from lucid-proposed
[21:09] <mathiaz> then I'd create a local branch to work on the sru itself
[21:10] <mathiaz> (in this use case I've used lucid-proposed as I know there is already an SRU in place)
[21:10] <mathiaz> (if not I'd use lp:ubuntu/lucid/eucalyptus - which would get the version of the src package that is in lucid)
[21:10] <mathiaz> then I'd create a local branch to work on the sru itself:
[21:11] <mathiaz> hm - wait - I usually check out the LP branch in a directory named after the release
[21:11] <mathiaz> so:
[21:11] <mathiaz> bzr co lp:ubuntu/lucid-proposed/eucalyptus lucid-proposed
[21:11] <ccheney> ok
[21:11] <mathiaz> that would create a sub-directory lucid-proposed/ that holds the relevant code
[21:12] <mathiaz> otherwise you'd end with eucalyptus as the directory name
[21:12] <mathiaz> then I'd create a local branch for preparing the SRU:
[21:12] <mathiaz> bzr branch lucid-proposed/ lucid-sru
[21:12] <mathiaz> cd lucid-sru/
[21:13] <mathiaz> and prepare the SRU in there (patches, changelog edition etc...)
[21:13] <ccheney> ok
[21:13] <mathiaz> in our use case you could just apply the debdiff
[21:13] <mathiaz> (lucid-sru:$ bzr patch < eucalyptus-sru.debdiff
[21:13] <mathiaz> )
[21:13] <mathiaz> once the sru is ready, do a local commit:
[21:13] <mathiaz> debcommit
[21:13] <mathiaz> and then push it to LP under your account:
[21:14] <mathiaz> bzr push lp:~ccheney/ubuntu/lucid-proposed/eucalyptus/lucid-sru
[21:15] <mathiaz> once the branch is pushed to LP you go over to the LP page
[21:15] <ccheney> does that need to be just lucid?
[21:15] <mathiaz> and you can propose for merging from the LP page
[21:15] <ccheney> it says no such distribution series
[21:16] <mathiaz> ccheney: ok
[21:16] <mathiaz> ccheney: so try lucid
[21:16] <Maletor> Can I access a DAAP share at home from work. I realize we need to forward ports, but doesn't it use bonjour? I would really like to be able to do this....
[21:16] <Maletor> SSHFS seems a little slow.......
[21:17] <mathiaz> ccheney: cool - your branch shows up: https://code.launchpad.net/ubuntu/lucid/+source/eucalyptus
[21:17] <ccheney> cool :)
[21:18] <mathiaz> ccheney: now if you go on the branch page: https://code.launchpad.net/~ccheney/ubuntu/lucid/eucalyptus/lucid-sru
[21:18] <mathiaz> ccheney: you can propose for merging
[21:18] <ccheney> i'm on the propose merge
[21:18] <ccheney> do i need to type the other url in as the target?
[21:18] <ccheney> it suggests lp:ubuntu/lucid/eucalyptus which seems wrong
[21:19] <mathiaz> ccheney: right - you wanna propose it to lucid-proposed
[21:19] <ccheney> pl
[21:19] <ccheney> er ok
[21:19] <mathiaz> ccheney: as this is what the branch is based on
[21:19] <ccheney> ok
[21:20] <ccheney> mathiaz, should i change the reviewer to ubuntu-server or something like that?
[21:21] <mathiaz> ccheney: yeah - put ubuntu-server as the reviewer
[21:21] <ccheney> ok
[21:21] <ccheney> ok done :)
[21:21] <ccheney> that was easy with help from you, thanks! :)
[21:22] <mathiaz> ccheney: cool - the merge is ready for review: https://code.launchpad.net/~ccheney/ubuntu/lucid/eucalyptus/lucid-sru/+merge/26839
[21:22] <mathiaz> ccheney: I'd suggest to subscribe a bunch of people to do the review as well
[21:22] <ccheney> ok
[21:22] <mathiaz> ccheney: I've just claimed the review
[21:23] <mathiaz> ccheney: https://code.launchpad.net/~ccheney/ubuntu/lucid/eucalyptus/lucid-sru/+merge/26839
[21:23] <ccheney> ok
[21:24] <mathiaz> ccheney: on the merge proposal page you can request another review
[21:24] <mathiaz> ccheney: I'd suggest to add Daviey as well
[21:24] <ccheney> ok
[21:25]  * ccheney gotta run to sell his house, bbiab
[21:25] <mathiaz> ccheney: as for the branch itself, I would subscribe kirkland and Daviey as well
[21:25] <ccheney> ok
[21:25] <mathiaz> ccheney: I'll do the later as you're running out
[21:25] <mathiaz> ccheney: good luck!
[21:28] <ccheney> i didn't see where to subscribe other people, i saw where they are subed
[21:28] <ccheney> but not how to add more
[21:28] <ccheney> looks like only me and james westby were subscribed by default
[21:29] <mathiaz> ccheney: on the branch page, there is a link that says subscribe someone else
[21:30] <ccheney> oh i was under the proposal page, oops
[21:30] <mathiaz> ccheney: right -it's a bit confusing
[21:30] <ccheney> thanks for the help :)
[21:30] <mathiaz> ccheney: https://code.launchpad.net/~ccheney/ubuntu/lucid/eucalyptus/lucid-sru
[21:31] <mathiaz> ccheney: ^^ that the bzr branch page
[21:31]  * ccheney leaving now, bbl
[21:31] <ccheney> yea
[21:31] <mathiaz> ccheney: that's all for now!
[21:31] <ccheney> ok
[21:37] <SpamapS> mathiaz: argh, I missed your message earlier
[21:37]  * SpamapS saves the log
[21:38] <mathiaz> SpamapS: that's ok - let me know if you have some questions
[22:24] <MrPicard> Hey guys
[22:24] <MrPicard> ive got a rackspace cloud server
[22:24] <MrPicard> im trying to configure ubuntu to work with the server and run ebox on top
[22:28] <bventura> hello!  I'm setting up a mail server with ubuntu and am wondering what the deal is with DKIM/DomainKeys.  I want to do everything I can to get my mail through to yahoo users.  Do I need to implement both DKIM AND DomainKeys?  Or is DKIM the new way of doing it, and DomainKeys obsolete?  From what I've read online so far it seems unclear as to which one to use or if I need both
[22:51] <ScottK> bventura: Yahoo! is rolling out DKIM and will retire DomainKeys.  I'd just worry about DKIM for something new.
[22:53] <LowValueTarget> im running an apt-mirror and im checking the logs every few days. I see something that says 959.3MiB can be freed.... run clean.sh
[22:53] <LowValueTarget> i run clean sh
[22:54] <LowValueTarget> and it says removing 260 unnecessary files..... done. but it doesnt seem to be removing them.
[22:54] <bventura> ScottK: thanks.
[22:54] <bventura> if you have a link to where you found that info on yahoo i'd appreciate it
[22:55] <guntbert> LowValueTarget: as what user are you running it? always the same one?
[22:59] <ScottK> bventura: I've worked on DKIM development.  In the community there I'm active in it's common knowledge.  I can't give you a specific link, although I'm sure it's public information.
[23:00] <bventura> ok
[23:03] <qman__> so, I had a NIC fail in a server, and I replaced it
[23:03] <qman__> and in order to make the new NIC be eth0, I removed the old line in /etc/udev/rules.d/70-persistent-net
[23:04] <qman__> but it's not giving the new one a name
[23:04] <qman__> how do I force detection of new cards?
[23:04] <qman__> it shows up in lspci
[23:05] <qman__> this is a hardy server, btw, 8.04.4
[23:23] <Scunizi> I've just turned on my server after several days of being off and it has no connectivity to the network.  ifup & down don't resolve the issue.. double checked wires and routers with no results.. before I go changing the nic what else can I test?
[23:24] <Scunizi> qman__: what does ifconfig say?
[23:24] <qman__> ifconfig doesn't see it
[23:24] <qman__> nothing in dmesg either
[23:25] <Scunizi> what does lspci call it?
[23:25] <qman__> Intel corporation 8-something gigabit adapter
[23:25] <qman__> 82574L
[23:26] <Scunizi> qman__: was the original nic built into the motherboard?
[23:26] <qman__> yes, nforce 4, using the forcedeth driver
[23:26] <qman__> I turned it off in the BIOS
[23:27] <Scunizi> try sudo lshw .. in the resulting output it will show the logical name for the card.  Even with the internal turned off it might be named eth1 or something
[23:29] <qman__> it's not getting one
[23:29] <qman__> lshw says it's unclaimed
[23:29] <qman__> that's the function I'm trying to find, how to have it detect and assign one
[23:29] <qman__> I know it normally does that automatically but for some reason it isn't
[23:30] <Scunizi> that's strange.. it should automatically discover and assign a logical name
[23:30] <Scunizi> maybe try a different pci slot?
[23:30] <qman__> it's PCI express, I could, but that'd be a waste of an x16 slot for an x1 card
[23:30] <qman__> and it is clearly getting detected
[23:31] <qman__> just not configured
[23:32] <qman__> hmm
[23:32] <qman__> google hints that maybe that card isn't supported with hardy's e1000 driver
[23:34] <Scunizi> qman__: possible.. you're still on hardy huh?  What's the content of /etc/network/interfaces?
[23:35] <qman__> just a normal single interface, eth0, with a static IP
[23:35] <Scunizi> although this is a slackware post it might shed some light on it. http://www.linuxforums.org/forum/slackware-linux-help/63049-detecting-network-card.html
[23:35] <qman__> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/513292
[23:39] <qman__> no workaround was posted, so it looks like the only fix is to compile the new driver from source
[23:39] <qman__> which should be an interesting trick without network connectivity
[23:39] <qman__> I don't think that machine has build-essential on it
[23:40] <Scunizi> ouch.. or look in packages.ubuntu.com for the driver?
[23:41] <mathiaz> ccheney: hey - what's the status of bug 566793?
[23:41] <hggdh> mathiaz: there?
[23:41] <mathiaz> ccheney: IIUC the fix will be reverted in the lucid SRU you're preparing?
[23:42] <mathiaz> hggdh: I hope so - otherwise something is fishy here...
[23:42] <hggdh> :-) I asked before I read the backlog...
[23:42] <mathiaz> hggdh: ah ok
[23:42] <mathiaz> hggdh: what are you up to/
[23:42] <mathiaz> hggdh: ?
[23:42] <hggdh> mathiaz: is there a compelling reason for the sources.list in the test rig to all point to 10.55.55.1?
[23:43] <hggdh> mathiaz: I just reinstalled topo2, after having 100% failures
[23:43] <mathiaz> hggdh: which install media are you using?
[23:43] <mathiaz> hggdh: the isos or the archive?
[23:43] <hggdh> a locally donwloaded iso...
[23:43] <hggdh> of course
[23:43] <mathiaz> hggdh: if you're using the isos it's normal that sources.list points 10.55.55.1 as this is where the isos are unpacked
[23:44] <hggdh> mathiaz: so I can reset them all, correct?
[23:44] <mathiaz> hggdh: sure
[23:44] <mathiaz> hggdh: they will just reinstall everything if you set them to on tamarind
[23:45] <hggdh> oops
[23:45] <hggdh> mathiaz: why?
[23:45] <mathiaz> hggdh: well - I'm not sure I fully understand what you're trying to do/ask
[23:46] <hggdh> mathiaz: I reinstalled from 10.04, ran some tests, and now I want to apply current fixes to 10.04
[23:46] <mathiaz> hggdh: so you're testing -proposed
[23:46] <mathiaz> hggdh: ?
[23:46] <hggdh> yes, that's the intention
[23:46] <mathiaz> hggdh: ah ok
[23:47] <hggdh> and starting to change the code to add in more tests
[23:47] <mathiaz> hggdh: so - you probably wanna install from the archive then
[23:47] <mathiaz> hggdh: instead from the iso
[23:47] <hggdh> k
[23:47] <mathiaz> hggdh: (although it should be the same)
[23:47] <hggdh> yes inideed. I will change the sources.list, and see what happens. Should just update
[23:47] <mathiaz> hggdh: if you're installing from the archive the sources.lists should not point to 10.55.55.1
[23:48] <hggdh> mathiaz: thank you
[23:49] <Daviey> mathiaz: i have seen a debdiff from ccheney that fixes bug 566793
[23:49] <mathiaz> Daviey: hm - fixes?
[23:49] <hggdh> it's in his branch
[23:49] <mathiaz> Daviey: my understanding was that the patch was actually reverted
[23:49] <Daviey> i haven't tested it, but he did have a debdiff ready IIRC
[23:49] <Daviey> hmm
[23:50] <mathiaz> Daviey: https://code.launchpad.net/~ccheney/ubuntu/lucid/eucalyptus/lucid-sru
[23:53] <mathiaz> Daviey: https://launchpad.net/ubuntu/+source/eucalyptus